Meta Launches Enterprise AI Platform Featuring Muse

Meta is making a major push into enterprise technology, launching Meta Enterprise Platform to sell AI agents, models and related services directly to businesses.

The company has recruited Chirantan “CJ” Desai, formerly CEO and president of MongoDB, to run the new operation as chief enterprise platform officer. He will report directly to Meta CEO Mark Zuckerberg.

The move is a major expansion for Meta, whose business has depended on advertising across Facebook and Instagram. Meta is spending more than $100 billion on AI infrastructure this year, creating pressure to generate revenue from that vast investment beyond its traditional advertising business.

“Meta is chasing volume in a market that’s rapidly filling up with AI productivity and development tools,” said Mitch Ashley, VP and Practice Lead at The Futurum Group. The problem, he noted, is that “ad creative generation and customer-facing bots are assistive work, exactly what buyers stop paying a premium for once every vendor offers it, so a bigger advertiser base doesn’t rebuild the pricing power Meta’s core ad business is losing.”

“Watch whether the roadmap moves past generation and bots toward outcome measurement and account-level governance, the work still scarce enough to price. Short of that, it’s more supply into a market that already has plenty.”

Turning Consumer Momentum Into an Enterprise Offering

Meta Enterprise Platform will initially include the company’s Muse agent, Meta Business Agent, Muse API and Muse Code. Zuckerberg said Meta plans to draw on its AI models, agents and computing infrastructure to build the enterprise operation.

Desai brings extensive enterprise software experience to these issues. Before leading MongoDB, he headed product and engineering at Cloudflare and spent nearly eight years at ServiceNow, including serving as president and chief operating officer.

But key details remain unannounced. Meta has not disclosed pricing or explained how the various products will be packaged. The company also has not detailed the administrative, security and contractual controls that large businesses typically require when deploying software across their operations.

This will be particularly important for Muse, Meta’s new personal AI agent. Muse can perform actions for users, including sending email and booking travel. The product has gained considerable early adoption, with Sensor Tower estimating more than 3.4 million downloads as of September 24. It also reached the top position in the U.S. Apple App Store and Google Play Store during its first two weeks.

The challenge for Meta is turning this consumer momentum into an enterprise-grade offering. AI agents become more useful when connected to email, applications, accounts and corporate data. On the other hand, those connections also create security concerns for IT departments.

Meta has built Muse with a dedicated virtual machine, separate credential storage and a monitoring system called Sentinel, which evaluates actions requested by the agent. Yet Meta’s current architecture does not technically block the company from accessing data inside a user’s virtual machine when required to operate or protect the service.

Meta plans a Confidential VM that would use encryption to prevent this access, but the company has not announced that capability as part of the current Muse product.

Meta enters this market with a mixed history in enterprise software. Its Workplace collaboration service grew to more than seven million paid subscribers by 2021, but Meta later discontinued the product, which closed in May 2026. Horizon Workrooms, its VR workplace collaboration application, shut down in February 2026.

This time, Meta has considerably more at stake. It is competing in an enterprise AI market that includes Microsoft, Google, Anthropic and OpenAI, while lower-cost models are adding pricing pressure across the sector.